My luck I had couple of older kernels, which work. The new kernel, probably somehow patched by VirtualBox fails to load. What fails is X-server, when it tries to start, here's the log:
[ 14.284] (II) LoadModule: "nvidia" [ 14.284] (II) Loading /usr/lib64/xorg/modules/drivers/nvidia_drv.so [ 14.285] (II) Module nvidia: vendor="NVIDIA Corporation" [ 14.285] compiled for 4.0.2, module version = 1.0.0 [ 14.285] Module class: X.Org Video Driver [ 14.294] (EE) NVIDIA: Failed to load the NVIDIA kernel module. Please check your [ 14.294] (EE) NVIDIA: system's kernel log for additional error messages. [ 14.294] (II) UnloadModule: "nvidia" [ 14.294] (II) Unloading nvidia [ 14.294] (EE) Failed to load module "nvidia" (module-specific error, 0) [ 14.294] (EE) No drivers available. [ 14.294] Fatal server error: [ 14.294] no screens found
Probably, I could just uninstall VirtualBox, then upgrade the kernel + NVidia drivers (oh god why...) and I'd get my laptop back to normal, however, this will mean no VirtualBox for me.
The question thus: is there any way to make friends between VirtualBox and NVidia? Perhaps I need some specific drivers for VirtualBox?
(I'm the author of the question, but due to some errors happening in account management, I no longer appear as such)
Below is what I could get from
dmesg but it's as terse as the note above:
[ 14.058801] nvidia: module license 'NVIDIA' taints kernel. [ 14.058807] Disabling lock debugging due to kernel taint [ 14.172337] input: HDA NVidia HDMI/DP,pcm=9 as /devices/pci0000:00/0000:00:01.0/0000:01:00.1/sound/card1/input13 [ 14.172636] input: HDA NVidia HDMI/DP,pcm=8 as /devices/pci0000:00/0000:00:01.0/0000:01:00.1/sound/card1/input14 [ 14.172888] input: HDA NVidia HDMI/DP,pcm=7 as /devices/pci0000:00/0000:00:01.0/0000:01:00.1/sound/card1/input15 [ 14.173166] input: HDA NVidia HDMI/DP,pcm=3 as /devices/pci0000:00/0000:00:01.0/0000:01:00.1/sound/card1/input16 [ 14.175252] vgaarb: device changed decodes: PCI:0000:01:00.0,olddecodes=io+mem,decodes=none:owns=io+mem [ 14.175527] NVRM: loading NVIDIA UNIX x86_64 Kernel Module 304.54 Sat Sep 29 00:05:49 PDT 2012
I will also look now into yum logs to see if there maybe was a record of what exactly did it do to the kernel / configuration.
Jul 26 00:53:18 Installed: kmod-VirtualBox-3.9.10-100.fc17.x86_64-4.2.12-1.fc17.13.x86_64 Jul 26 00:53:19 Installed: kmod-VirtualBox-4.2.12-1.fc17.13.x86_64 Jul 26 00:53:25 Installed: VirtualBox-4.2.12-1.fc17.x86_64 Jul 26 00:53:25 Erased: 1:kmod-nvidia-3.5.3-1.fc17.x86_64-304.51-1.fc17.x86_64
OK, so, apparently,
yum decided it needs to erase the NVidia kernel module... Is my guess correct then, that these two are incompatible? And if so, there is no way they will work together? What would be an alternative, if any to VirtualBox?
Sorry, I cannot comment yet. Just to make it clear: the driver is for the host system, not the guest. I actually took courage to just remove all things related to NVidia, reinstall them and VirtualBox started allright. What I did would be something along the lines of:
# yum update kernel* selinux-policy* # yum list installed | grep nvidia | yum remove # reboot # yum install akmod-nvidia nvidia-settings nvidia-xconfig
I think I had CUDA drivers (used for development for NVidia paralleling technology), but I didn't use them anyway.